The integration of artificial intelligence (AI) in peer review is gaining traction, with many experts believing it's the future of scientific publishing. The sheer volume of research papers being submitted is overwhelming, with an estimated 10 million peer reviews needed this year alone. To address this challenge, AI can assist editors in deciding which articles to send for external peer review, making the process more efficient and effective.
AI can handle large workloads without compromising quality, reducing reviewer burnout and delays in the publishing process. Additionally, AI models can be trained to disregard author identities and affiliations, reducing biases in the review process. AI can also ensure that reporting guidelines are followed, and adherence to these guidelines is checked consistently.
However, there are concerns about the limitations and potential risks of AI in peer review. AI tools can perpetuate biases present in the training data, and their decision-making processes may not be transparent. Moreover, AI tools can potentially leak sensitive information or compromise author anonymity. Over-reliance on AI may also lead to the devaluation of human expertise and judgment in the peer review process.
To address these concerns, it's essential to develop guidelines and policies for the use of AI in peer review, ensuring transparency, accountability, and the complementary use of human expertise.
